Delta DPS400-MB, pwr_ok stays low

[quicknick]

Howdy folks,

I recently got this PSU to replace the 250W Delta unit in my "Retro-PC workhorse" - the rig that I'm using for testing most of my retro stuff.

https://www.badcaps.net/forum/attach...1&d=1648432776

https://www.badcaps.net/forum/attach...1&d=1648432776

https://www.badcaps.net/forum/attach...1&d=1648432776

Cleaned it good (wasn't too dirty to begin with) then proceeded to test it with a minimal load (two old hard drives). Everything looked OK, all voltages in spec except the pwr_ok signal that fluctuates somewhere between 0 and 0.4 volts.

Despite this, I used it to power the "workhorse" and surprisingly it started and ran as usual (pwr_ok still low), but I'm wary of using it like this on a test system so I reverted to the old 250W supply.

Took it apart again, desoldered the small daughterboard/control board to check for problems. At first I thought it was soaked in electrolyte from supposedly bad capacitors, but this proved to be only an excess of flux from manufacturing, which was easy to clean with iso alcohol.

One of the small elkos on the board (nichicon PW 10uF/25V) seemed to have a high-ish ESR (a bit over 4 ohm) so I replaced it with a 10uF/35V from the same series (around 1 ohm I think). However, after putting it all together it didn't change anything. Voltages still in spec, pwr_ok missing.

The three chips on the small board are two LM339 (one DIP, one SOIC) and a TSM1002DS which I suspect could be a custom Delta chip. All capacitors are from reputable brands (NCC primary, mostly Rubycon on the main board, Nichicon on the daughterboard). Lacking a schematic (and experience) I'm at a loss, so any advice/idea/hint will be greatly appreciated.

Thanks!

[PeteS in CA]

Re: Delta DPS400-MB, pwr_ok stays low

You may be able to find the pin-out and a schematic for that custom Delta ASIC. It was “DNA1002” when I was at Delta in the late 90s. Look to see if the PG output needs a pull-up resistor, and if there is one on the PCB. If there is, check to see if it is open or has a bad solder joint.

[Topcat]

Re: Delta DPS400-MB, pwr_ok stays low

Just for future reference, please attach images to your post and not use 3rd party / off-site image hosts.

I fixed it this go-round.

[quicknick]

Re: Delta DPS400-MB, pwr_ok stays low

It was “DNA1002” when I was at Delta in the late 90s.

Thanks for the hint! I searched around, found some schematics and also this thread describing a very similar problem.
I'll investigate further in the following days, checking the small caps on the mainboard and the voltages around the chip, maybe replacing the two LM339.

@Topcat - thanks, I guess I'm too old-fashioned (or maybe just old) - I was trying to save this board some bandwidth

[kaboom]

Re: Delta DPS400-MB, pwr_ok stays low

There should be a "pullup" resistor from +5 to PGO, as the latter is typically an "open collector" output.
If this is open, PG will never go high, no matter what's going on with the opamps/comparators.
For testing, connect anything from 1k-4.7k from +5 to PGO.
Easy to do at the ATX connector- connect the resistor from red to gray.
If PGO behaves normally, you're on the right track...
You have to trace PGO from the gray wire back to wherever it originates.

[quicknick]

Re: Delta DPS400-MB, pwr_ok stays low

Thanks for chiming in, every bit of advice got me closer to solving the problem.

First I tried the external resistor trick sugested by kaboom, and it seemed to indeed solve the problem. So I opened the PSU again and inspected the small board. Traced the gray wire and it only connected to a single point inside the PSU, that being pin 3 of the '1002' chip.

There are two pair of pads in the area, one for pulling P.G. to +5V, the other for pulling it to GND. None was populated - the PSU came like this from the factory. The unit has an IBM part number, so maybe this is some trick requested by them to make it less usable with other systems?

According to the schematics found online, I added a 1K to the "pull up" pads and after this the PSU behaves as expected. To make sure I didn't simply pull up a dead circuit i quickly wired some LEDs+resistors to the main rails and also the PG wire, and as expected there's a small delay (about 0.2s) for the PG compared to the rails.

Thanks again for everything, I now consider this solved!

[PeteS in CA]

Re: Delta DPS400-MB, pwr_ok stays low

I'm going to guess that the pull-up resistor for the Power Good signal is on the IBM MB, and that may be the case with your "Retro-PC workhorse".

[kaboom]

Re: Delta DPS400-MB, pwr_ok stays low

Thanks for chiming in, every bit of advice got me closer to solving the problem.

First I tried the external resistor trick sugested by kaboom, and it seemed to indeed solve the problem. So I opened the PSU again and inspected the small board. Traced the gray wire and it only connected to a single point inside the PSU, that being pin 3 of the '1002' chip.

There are two pair of pads in the area, one for pulling P.G. to +5V, the other for pulling it to GND. None was populated - the PSU came like this from the factory. The unit has an IBM part number, so maybe this is some trick requested by them to make it less usable with other systems?

According to the schematics found online, I added a 1K to the "pull up" pads and after this the PSU behaves as expected. To make sure I didn't simply pull up a dead circuit i quickly wired some LEDs+resistors to the main rails and also the PG wire, and as expected there's a small delay (about 0.2s) for the PG compared to the rails.

Thanks again for everything, I now consider this solved!

Excellent, another one saved/fixed!


IIRC, there were some AT-era Compaqs that had the PG circuitry on the MB.
I found this out when building a "retro PC" with a PS from one of these, several years ago.
There was no PG line on either of the AT plugs, so I built a PG comparator and mounted it in the PS case.